The broad purpose of the project is to lay the basis for the development of a behavioral assay of intellective deficit in early infancy. Normal, premature, and Down's syndrome infants are being compared in terms of their ability to code relational features of stimuli under the assumption that the processing of patterned information is fumdamental to intelligent behavior. Habituation-recovery procedures are being used to assess sensitivity to invariant spatial, temporal and social relationships between 48 and 78 weeks conceptional age. To evaluate predictive validity, performance in infancy will be related to cognitive and linguistic outcomes in childhood, and the extent of relationship will be compared with that of standard infant psychometric instruments.